Added PR_SET_PROCTITLE_AREA option for prctl()
Subsystems (5)
PATCH v4 1/1 Added PR_SET_PROCTITLE_AREA option for prctl()
ChangeLog
v3 -> v4
- Use mutex instead seq_lock as akpm requested.
========================================
From: Timo Sirainen <[email protected]>
Currently glibc2 doesn't have setproctitle(3), so several userland daemons attempt to emulate it by doing some brutal stack modifications. This works most of the time, but it has problems. For example:
% ps -ef |grep avahi-daemon
avahi 1679 1 0 09:20 ? 00:00:00 avahi-daemon: running [kosadesk.local]
# cat /proc/1679/cmdline avahi-daemon: running [kosadesk.local]
This looks good, but the process has also overwritten its environment area and made the environ file useless:
# cat /proc/1679/environ adesk.local]
Another problem is that the process title length is limited by the size of the environment. Security conscious people try to avoid potential information leaks by clearing most of the environment before running a daemon:
# env - MINIMUM_NEEDED_VAR=foo /path/to/daemon
The resulting environment size may be too small to fit the wanted process titles.
This patch makes it possible for userspace to implement setproctitle() cleanly. It adds a new PR_SET_PROCTITLE_AREA option for prctl(), which updates task's mm_struct->arg_start and arg_end to the given area.
test_setproctitle.c
================================================
#include <string.h>
#include <stdlib.h>
#include <unistd.h>
#include <stdio.h>
#include <sys/prctl.h>
#define ERR(str) (perror(str), exit(1))
void settitle(char* title){
int err;
err = prctl(34, title, strlen(title)+1);
if (err < 0)
ERR("prctl ");}
void main(void){
long i;
char buf[1024];
for (i = 0; i < 10000000000LL; i++){
sprintf(buf, "loooooooooooooooooooooooong string %d",i);
settitle(buf);
} }
==================================================
--- fs/proc/base.c | 31 ++++++++++++++++++++++--------- include/linux/mm_types.h | 2 ++ include/linux/prctl.h | 3 +++ kernel/fork.c | 1 + kernel/sys.c | 22 ++++++++++++++++++++++ 5 files changed, 50 insertions(+), 9 deletions(-)
fs/proc/base.c +22 -9
@@ -255,32 +255,45 @@ static int proc_pid_cmdline(struct task_struct *task, char * buffer)int res =0 ;unsigned int len;struct mm_struct *mm =get_task_mm (task); +if (!mm) goto out; + +/* The process was not constructed yet? */ if (!mm->arg_end) goto out_mm;/* Shh! No looking before we're done */ - len = mm->arg_end - mm->arg_start; - +mutex_lock (&mm->arg_lock); + len = mm->arg_end - mm->arg_start;if (len >PAGE_SIZE ) len =PAGE_SIZE ; - + res =access_process_vm (task, mm->arg_start, buffer, len,0 ); +if (mm->arg_end != mm->env_start) +/* prctl(PR_SET_PROCTITLE_AREA) used */ + goto out_unlock; -// If the nul at the end of args has been overwritten, then -// assume application is using setproctitle(3). +/* +* If the nul at the end of args has been overwritten, then assume +* application is using sendmail's SPT_REUSEARGV style argv override. +*/ if (res >0 && buffer[res-1 ] !='\0' && len <PAGE_SIZE ) { len =strnlen (buffer, res); -if (len < res) { - res = len; - }else { +if (len < res) + res = len; +else { len = mm->env_end - mm->env_start;if (len >PAGE_SIZE - res) len =PAGE_SIZE - res; - res +=access_process_vm (task, mm->env_start, buffer+res, len,0 ); + res +=access_process_vm (task, mm->env_start, + buffer+res, len,0 ); res =strnlen (buffer, res); } } + +out_unlock: +mutex_unlock (&mm->arg_lock); + out_mm:mmput (mm); out:
include/linux/mm_types.h +2 -0
@@ -12,6 +12,7 @@#include <linux/completion.h> #include <linux/cpumask.h> #include <linux/page-debug-flags.h> +#include <linux/mutex.h> #include <asm/page.h> #include <asm/mmu.h>
@@ -236,6 +237,7 @@ struct mm_struct {unsigned long stack_vm, reserved_vm, def_flags, nr_ptes;unsigned long start_code, end_code, start_data, end_data;unsigned long start_brk, brk, start_stack; +struct mutex arg_lock;unsigned long arg_start, arg_end, env_start, env_end;unsigned long saved_auxv[AT_VECTOR_SIZE ];/* for /proc/PID/auxv */
include/linux/prctl.h +3 -0
@@ -90,4 +90,7 @@#define PR_MCE_KILL 33 +/* Set process title memory area for setproctitle() */ +#define PR_SET_PROCTITLE_AREA 34 +#endif /* _LINUX_PRCTL_H */
kernel/fork.c +1 -0
@@ -459,6 +459,7 @@ static struct mm_struct * mm_init(struct mm_struct * mm, struct task_struct *p) mm->cached_hole_size = ~0UL ;mm_init_aio (mm);mm_init_owner (mm, p); +mutex_init (&mm->arg_lock);if (likely (!mm_alloc_pgd (mm))) { mm->def_flags =0 ;
kernel/sys.c +22 -0
@@ -1564,6 +1564,28 @@ SYSCALL_DEFINE5(prctl, int, option, unsigned long, arg2, unsigned long, arg3, error =0 ;break ; +case PR_SET_PROCTITLE_AREA : { +struct mm_struct *mm = current->mm; +unsigned long addr = arg2; +unsigned long len = arg3; +unsigned long end = arg2 + arg3; + +if (len >PAGE_SIZE ) +return -EINVAL ; + +if (addr >= end) +return -EINVAL ; + +if (!access_ok (VERIFY_READ , addr, len)) +return -EFAULT ; + +mutex_lock (&mm->arg_lock); + mm->arg_start = addr; + mm->arg_end = addr + len; +mutex_unlock (&mm->arg_lock); + +return 0 ; + }default : error = -EINVAL ;break ;
Signature · 1 line
1.6.2.5
Quoted 8 lines · KOSAKI Motohiro
ChangeLog
v3 -> v4
- Use mutex insteadseq_lockas akpm requested.========================================
From: Timo Sirainen <[email protected]>
Please see my comments below.
Quoted 189 lines
Currently glibc2 doesn't have setproctitle(3), so several userland daemons attempt to emulate it by doing some brutal stack modifications. This works most of the time, but it has problems. For example:
% ps -ef |grep avahi-daemon
avahi 1679 1 0 09:20 ? 00:00:00 avahi-daemon: running [kosadesk.local]# cat /proc/1679/cmdline avahi-daemon: running [kosadesk.local]This looks good, but the process has also overwritten its environment area and made the environ file useless:
# cat /proc/1679/environ adesk.local]Another problem is that the process title length is limited by the size of the environment. Security conscious people try to avoid potential information leaks by clearing most of the environment before running a daemon:
# env - MINIMUM_NEEDED_VAR=foo /path/to/daemonThe resulting environment size may be too small to fit the wanted process titles.
This patch makes it possible for userspace to implement
setproctitle()cleanly. It adds a newPR_SET_PROCTITLE_AREAoption forprctl(), which updates task'smm_struct->arg_startandarg_endto the given area.
test_setproctitle.c
================================================
#include <string.h>
#include <stdlib.h>
#include <unistd.h>
#include <stdio.h>
#include <sys/prctl.h>#define ERR(str) (perror(str), exit(1))
void settitle(char* title){ int err; err = prctl(34, title, strlen(title)+1); if (err < 0) ERR("prctl "); }void main(void){ long i; char buf[1024]; for (i = 0; i < 10000000000LL; i++){ sprintf(buf, "loooooooooooooooooooooooong string %d",i); settitle(buf); } }==================================================
Cc Bryan Donlan <[email protected]>Cc Ulrich Drepper <[email protected]>Signed-off-by KOSAKI Motohiro <[email protected]>Signed-off-by Timo Sirainen <[email protected]>---
fs/proc/base.c | 31 ++++++++++++++++++++++--------- include/linux/mm_types.h | 2 ++ include/linux/prctl.h | 3 +++ kernel/fork.c | 1 + kernel/sys.c | 22 ++++++++++++++++++++++ 5 files changed, 50 insertions(+), 9 deletions(-)fs/proc/base.c +22
-9@@ -255,32 +255,45 @@ static int proc_pid_cmdline(struct task_struct *task, char * buffer)int res =0 ;unsigned int len;struct mm_struct *mm =get_task_mm (task); +if (!mm) goto out; + +/* The process was not constructed yet? */ if (!mm->arg_end) goto out_mm;/* Shh! No looking before we're done */ - len = mm->arg_end - mm->arg_start; - +mutex_lock (&mm->arg_lock); + len = mm->arg_end - mm->arg_start;if (len >PAGE_SIZE ) len =PAGE_SIZE ; - + res =access_process_vm (task, mm->arg_start, buffer, len,0 ); +if (mm->arg_end != mm->env_start) +/* prctl(PR_SET_PROCTITLE_AREA) used */ + goto out_unlock; -// If the nul at the end of args has been overwritten, then -// assume application is using setproctitle(3). +/* +* If the nul at the end of args has been overwritten, then assume +* application is using sendmail's SPT_REUSEARGV style argv override. +*/ if (res >0 && buffer[res-1 ] !='\0' && len <PAGE_SIZE ) { len =strnlen (buffer, res); -if (len < res) { - res = len; - }else { +if (len < res) + res = len; +else { len = mm->env_end - mm->env_start;if (len >PAGE_SIZE - res) len =PAGE_SIZE - res; - res +=access_process_vm (task, mm->env_start, buffer+res, len,0 ); + res +=access_process_vm (task, mm->env_start, + buffer+res, len,0 ); res =strnlen (buffer, res); } } + +out_unlock: +mutex_unlock (&mm->arg_lock); +
out_mm:mmput(mm);out:
include/linux/mm_types.h +2
-0@@ -12,6 +12,7 @@#include <linux/completion.h> #include <linux/cpumask.h> #include <linux/page-debug-flags.h> +#include <linux/mutex.h> #include <asm/page.h> #include <asm/mmu.h> @@ -236,6 +237,7 @@ struct mm_struct {unsigned long stack_vm, reserved_vm, def_flags, nr_ptes;unsigned long start_code, end_code, start_data, end_data;unsigned long start_brk, brk, start_stack; +struct mutex arg_lock;unsigned long arg_start, arg_end, env_start, env_end;unsigned long saved_auxv[AT_VECTOR_SIZE ];/* for /proc/PID/auxv */ include/linux/prctl.h +3
-0@@ -90,4 +90,7 @@#define PR_MCE_KILL 33 +/* Set process title memory area for setproctitle() */ +#define PR_SET_PROCTITLE_AREA 34 +#endif /* _LINUX_PRCTL_H */ kernel/fork.c +1
-0@@ -459,6 +459,7 @@ static struct mm_struct * mm_init(struct mm_struct * mm, struct task_struct *p) mm->cached_hole_size = ~0UL ;mm_init_aio (mm);mm_init_owner (mm, p); +mutex_init (&mm->arg_lock);if (likely (!mm_alloc_pgd (mm))) { mm->def_flags =0 ;kernel/sys.c +7
-0@@ -1564,6 +1564,28 @@ SYSCALL_DEFINE5(prctl, int, option, unsigned long, arg2, unsigned long, arg3, error =0 ;break ; +case PR_SET_PROCTITLE_AREA : { +struct mm_struct *mm = current->mm; +unsigned long addr = arg2; +unsigned long len = arg3; +unsigned long end = arg2 + arg3; + +if (len >PAGE_SIZE )
kernel/sys.c SYSCALL_DEFINE5(prctl, int, option, unsigned long, arg2, unsigned long, arg3, · line 1583
Quoted 10 lines
kernel/sys.c +10
-0+return -EINVAL ; + +if (addr >= end) +return -EINVAL ; + +if (!access_ok (VERIFY_READ , addr, len)) +return -EFAULT ; + +mutex_lock (&mm->arg_lock); + mm->arg_start = addr;
Is this safe? You're assigning a user-space pointer to kernel space... Don't we need copy_from_user()?
kernel/sys.c SYSCALL_DEFINE5(prctl, int, option, unsigned long, arg2, unsigned long, arg3, · line 1584
Quoted 1 line
kernel/sys.c +1
-0+ mm->arg_end = addr + len;
Since you already have 'end', no need to caculate this again. :)
Quoted 7 lines
kernel/sys.c +4
-0+mutex_unlock (&mm->arg_lock); + +return 0 ; + }default : error = -EINVAL ;break ;
Signature · 1 line
Live like a child, think like the god.
Quoted 12 lines
Quoted 12 lines
kernel/sys.c +12
-0+case PR_SET_PROCTITLE_AREA : { +struct mm_struct *mm = current->mm; +unsigned long addr = arg2; +unsigned long len = arg3; +unsigned long end = arg2 + arg3; + +if (len >PAGE_SIZE ) +return -EINVAL ; + +if (addr >= end) +return -EINVAL ; +
Quoted 8 lines
kernel/sys.c SYSCALL_DEFINE5(prctl, int, option, unsigned long, arg2, unsigned long, arg3, · line 1583
Quoted 5 lines
kernel/sys.c +5
-0+if (!access_ok (VERIFY_READ , addr, len)) +return -EFAULT ; + +mutex_lock (&mm->arg_lock); + mm->arg_start = addr;Is this safe? You're assigning a user-space pointer to kernel space... Don't we need
copy_from_user()?
mm->arg_start, arg_end are defined so.
Please see current implementation.
Quoted 3 lines
kernel/sys.c SYSCALL_DEFINE5(prctl, int, option, unsigned long, arg2, unsigned long, arg3, · line 1584
Quoted 1 line
kernel/sys.c +1
-0+ mm->arg_end = addr + len;Since you already have 'end', no need to caculate this again. :)
Good catch :)
PATCH v5 1/1 Added PR_SET_PROCTITLE_AREA option for prctl()
Quoted 25 lines
Quoted 12 lines
Quoted 12 lines
(quoted diff) +12
-0+ case PR_SET_PROCTITLE_AREA: { + struct mm_struct *mm = current->mm; + unsigned long addr = arg2; + unsigned long len = arg3; + unsigned long end = arg2 + arg3; + + if (len > PAGE_SIZE) + return -EINVAL; + + if (addr >= end) + return -EINVAL; +Quoted 8 lines
Review
Quoted 5 lines
(quoted diff) +5
-0+ if (!access_ok(VERIFY_READ, addr, len)) + return -EFAULT; + + mutex_lock(&mm->arg_lock); + mm->arg_start = addr;Is this safe? You're assigning a user-space pointer to kernel space... Don't we need
copy_from_user()?
mm->arg_start,arg_endare defined so.
Please see current implementation.
Quoted 7 lines
Quoted 3 lines
Review
Quoted 1 line
(quoted diff) +1
-0+ mm->arg_end = addr + len;Since you already have 'end', no need to caculate this again. :)
Good catch :)
Fixed.
ChangeLog
v4 -> v5
- nit: kill duplicate calculation in prctl()
v3 -> v4
- Use mutex instead seq_lock.
========================================
Subject: [PATCH v5] Added PR_SET_PROCTITLE_AREA option for prctl()
From: Timo Sirainen <[email protected]>
Currently glibc2 doesn't have setproctitle(3), so several userland daemons attempt to emulate it by doing some brutal stack modifications. This works most of the time, but it has problems. For example:
% ps -ef |grep avahi-daemon
avahi 1679 1 0 09:20 ? 00:00:00 avahi-daemon: running [kosadesk.local]
# cat /proc/1679/cmdline avahi-daemon: running [kosadesk.local]
This looks good, but the process has also overwritten its environment area and made the environ file useless:
# cat /proc/1679/environ adesk.local]
Another problem is that the process title length is limited by the size of the environment. Security conscious people try to avoid potential information leaks by clearing most of the environment before running a daemon:
# env - MINIMUM_NEEDED_VAR=foo /path/to/daemon
The resulting environment size may be too small to fit the wanted process titles.
This patch makes it possible for userspace to implement setproctitle() cleanly. It adds a new PR_SET_PROCTITLE_AREA option for prctl(), which updates task's mm_struct->arg_start and arg_end to the given area.
test_setproctitle.c
================================================
#include <string.h>
#include <stdlib.h>
#include <unistd.h>
#include <stdio.h>
#include <sys/prctl.h>
#define ERR(str) (perror(str), exit(1))
void settitle(char* title){
int err;
err = prctl(34, title, strlen(title)+1);
if (err < 0)
ERR("prctl ");}
void main(void){
long i;
char buf[1024];
for (i = 0; i < 10000000000LL; i++){
sprintf(buf, "loooooooooooooooooooooooong string %d",i);
settitle(buf);
} }
==================================================
--- fs/proc/base.c | 31 ++++++++++++++++++++++--------- include/linux/mm_types.h | 2 ++ include/linux/prctl.h | 3 +++ kernel/fork.c | 1 + kernel/sys.c | 22 ++++++++++++++++++++++ 5 files changed, 50 insertions(+), 9 deletions(-)
fs/proc/base.c +22 -9
@@ -255,32 +255,45 @@ static int proc_pid_cmdline(struct task_struct *task, char * buffer)int res =0 ;unsigned int len;struct mm_struct *mm =get_task_mm (task); +if (!mm) goto out; + +/* The process was not constructed yet? */ if (!mm->arg_end) goto out_mm;/* Shh! No looking before we're done */ - len = mm->arg_end - mm->arg_start; - +mutex_lock (&mm->arg_lock); + len = mm->arg_end - mm->arg_start;if (len >PAGE_SIZE ) len =PAGE_SIZE ; - + res =access_process_vm (task, mm->arg_start, buffer, len,0 ); +if (mm->arg_end != mm->env_start) +/* prctl(PR_SET_PROCTITLE_AREA) used */ + goto out_unlock; -// If the nul at the end of args has been overwritten, then -// assume application is using setproctitle(3). +/* +* If the nul at the end of args has been overwritten, then assume +* application is using sendmail's SPT_REUSEARGV style argv override. +*/ if (res >0 && buffer[res-1 ] !='\0' && len <PAGE_SIZE ) { len =strnlen (buffer, res); -if (len < res) { - res = len; - }else { +if (len < res) + res = len; +else { len = mm->env_end - mm->env_start;if (len >PAGE_SIZE - res) len =PAGE_SIZE - res; - res +=access_process_vm (task, mm->env_start, buffer+res, len,0 ); + res +=access_process_vm (task, mm->env_start, + buffer+res, len,0 ); res =strnlen (buffer, res); } } + +out_unlock: +mutex_unlock (&mm->arg_lock); + out_mm:mmput (mm); out:
include/linux/mm_types.h +2 -0
@@ -12,6 +12,7 @@#include <linux/completion.h> #include <linux/cpumask.h> #include <linux/page-debug-flags.h> +#include <linux/mutex.h> #include <asm/page.h> #include <asm/mmu.h>
@@ -236,6 +237,7 @@ struct mm_struct {unsigned long stack_vm, reserved_vm, def_flags, nr_ptes;unsigned long start_code, end_code, start_data, end_data;unsigned long start_brk, brk, start_stack; +struct mutex arg_lock;unsigned long arg_start, arg_end, env_start, env_end;unsigned long saved_auxv[AT_VECTOR_SIZE ];/* for /proc/PID/auxv */
include/linux/prctl.h +3 -0
@@ -90,4 +90,7 @@#define PR_MCE_KILL 33 +/* Set process title memory area for setproctitle() */ +#define PR_SET_PROCTITLE_AREA 34 +#endif /* _LINUX_PRCTL_H */
kernel/fork.c +1 -0
@@ -459,6 +459,7 @@ static struct mm_struct * mm_init(struct mm_struct * mm, struct task_struct *p) mm->cached_hole_size = ~0UL ;mm_init_aio (mm);mm_init_owner (mm, p); +mutex_init (&mm->arg_lock);if (likely (!mm_alloc_pgd (mm))) { mm->def_flags =0 ;
kernel/sys.c +22 -0
@@ -1564,6 +1564,28 @@ SYSCALL_DEFINE5(prctl, int, option, unsigned long, arg2, unsigned long, arg3, error =0 ;break ; +case PR_SET_PROCTITLE_AREA : { +struct mm_struct *mm = current->mm; +unsigned long addr = arg2; +unsigned long len = arg3; +unsigned long end = arg2 + arg3; + +if (len >PAGE_SIZE ) +return -EINVAL ; + +if (addr >= end) +return -EINVAL ; + +if (!access_ok (VERIFY_READ , addr, len)) +return -EFAULT ; + +mutex_lock (&mm->arg_lock); + mm->arg_start = addr; + mm->arg_end = end; +mutex_unlock (&mm->arg_lock); + +return 0 ; + }default : error = -EINVAL ;break ;
Signature · 1 line
1.6.2.5
Quoted 21 lines · KOSAKI Motohiro
Quoted 21 lines
Quoted 20 lines
Quoted 7 lines
kernel/sys.c +7
-0+case PR_SET_PROCTITLE_AREA : { +struct mm_struct *mm = current->mm; +unsigned long addr = arg2; +unsigned long len = arg3; +unsigned long end = arg2 + arg3; + +if (len >PAGE_SIZE )kernel/sys.c SYSCALL_DEFINE5(prctl, int, option, unsigned long, arg2, unsigned long, arg3, · line 1583
Quoted 10 lines
kernel/sys.c +10
-0+return -EINVAL ; + +if (addr >= end) +return -EINVAL ; + +if (!access_ok (VERIFY_READ , addr, len)) +return -EFAULT ; + +mutex_lock (&mm->arg_lock); + mm->arg_start = addr;Is this safe? You're assigning a user-space pointer to kernel space... Don't we need
copy_from_user()?
Quoted 2 lines
Quoted 2 lines
mm->arg_start,arg_endare defined so.
Please see current implementation.
Hmm, yeah.
Quoted 80 lines
Quoted 9 lines
Quoted 5 lines
Review
Quoted 3 lines
(quoted diff) +1
-0+ mm->arg_end = addr + len;Since you already have 'end', no need to caculate this again. :)
Good catch :)
Fixed.
ChangeLog
v4 -> v5
- nit: kill duplicate calculation inprctl()
v3 -> v4
- Use mutex insteadseq_lock.========================================
Subject: [PATCH v5] Added
PR_SET_PROCTITLE_AREAoption forprctl()
From: Timo Sirainen <[email protected]>Currently glibc2 doesn't have setproctitle(3), so several userland daemons attempt to emulate it by doing some brutal stack modifications. This works most of the time, but it has problems. For example:
% ps -ef |grep avahi-daemon
avahi 1679 1 0 09:20 ? 00:00:00 avahi-daemon: running [kosadesk.local]# cat /proc/1679/cmdline avahi-daemon: running [kosadesk.local]This looks good, but the process has also overwritten its environment area and made the environ file useless:
# cat /proc/1679/environ adesk.local]Another problem is that the process title length is limited by the size of the environment. Security conscious people try to avoid potential information leaks by clearing most of the environment before running a daemon:
# env - MINIMUM_NEEDED_VAR=foo /path/to/daemonThe resulting environment size may be too small to fit the wanted process titles.
This patch makes it possible for userspace to implement
setproctitle()cleanly. It adds a newPR_SET_PROCTITLE_AREAoption forprctl(), which updates task'smm_struct->arg_startandarg_endto the given area.
test_setproctitle.c
================================================
#include <string.h>
#include <stdlib.h>
#include <unistd.h>
#include <stdio.h>
#include <sys/prctl.h>#define ERR(str) (perror(str), exit(1))
void settitle(char* title){ int err; err = prctl(34, title, strlen(title)+1); if (err < 0) ERR("prctl "); }void main(void){ long i; char buf[1024]; for (i = 0; i < 10000000000LL; i++){ sprintf(buf, "loooooooooooooooooooooooong string %d",i); settitle(buf); } }
Quoted 6 lines
==================================================
Cc Bryan Donlan <[email protected]>Cc Ulrich Drepper <[email protected]>Signed-off-by KOSAKI Motohiro <[email protected]>Signed-off-by Timo Sirainen <[email protected]>
This version looks good,
Thanks!
Quoted 150 lines
---
fs/proc/base.c | 31 ++++++++++++++++++++++--------- include/linux/mm_types.h | 2 ++ include/linux/prctl.h | 3 +++ kernel/fork.c | 1 + kernel/sys.c | 22 ++++++++++++++++++++++ 5 files changed, 50 insertions(+), 9 deletions(-)fs/proc/base.c +22
-9@@ -255,32 +255,45 @@ static int proc_pid_cmdline(struct task_struct *task, char * buffer)int res =0 ;unsigned int len;struct mm_struct *mm =get_task_mm (task); +if (!mm) goto out; + +/* The process was not constructed yet? */ if (!mm->arg_end) goto out_mm;/* Shh! No looking before we're done */ - len = mm->arg_end - mm->arg_start; - +mutex_lock (&mm->arg_lock); + len = mm->arg_end - mm->arg_start;if (len >PAGE_SIZE ) len =PAGE_SIZE ; - + res =access_process_vm (task, mm->arg_start, buffer, len,0 ); +if (mm->arg_end != mm->env_start) +/* prctl(PR_SET_PROCTITLE_AREA) used */ + goto out_unlock; -// If the nul at the end of args has been overwritten, then -// assume application is using setproctitle(3). +/* +* If the nul at the end of args has been overwritten, then assume +* application is using sendmail's SPT_REUSEARGV style argv override. +*/ if (res >0 && buffer[res-1 ] !='\0' && len <PAGE_SIZE ) { len =strnlen (buffer, res); -if (len < res) { - res = len; - }else { +if (len < res) + res = len; +else { len = mm->env_end - mm->env_start;if (len >PAGE_SIZE - res) len =PAGE_SIZE - res; - res +=access_process_vm (task, mm->env_start, buffer+res, len,0 ); + res +=access_process_vm (task, mm->env_start, + buffer+res, len,0 ); res =strnlen (buffer, res); } } + +out_unlock: +mutex_unlock (&mm->arg_lock); +
out_mm:mmput(mm);out:
include/linux/mm_types.h +2
-0@@ -12,6 +12,7 @@#include <linux/completion.h> #include <linux/cpumask.h> #include <linux/page-debug-flags.h> +#include <linux/mutex.h> #include <asm/page.h> #include <asm/mmu.h> @@ -236,6 +237,7 @@ struct mm_struct {unsigned long stack_vm, reserved_vm, def_flags, nr_ptes;unsigned long start_code, end_code, start_data, end_data;unsigned long start_brk, brk, start_stack; +struct mutex arg_lock;unsigned long arg_start, arg_end, env_start, env_end;unsigned long saved_auxv[AT_VECTOR_SIZE ];/* for /proc/PID/auxv */ include/linux/prctl.h +3
-0@@ -90,4 +90,7 @@#define PR_MCE_KILL 33 +/* Set process title memory area for setproctitle() */ +#define PR_SET_PROCTITLE_AREA 34 +#endif /* _LINUX_PRCTL_H */ kernel/fork.c +1
-0@@ -459,6 +459,7 @@ static struct mm_struct * mm_init(struct mm_struct * mm, struct task_struct *p) mm->cached_hole_size = ~0UL ;mm_init_aio (mm);mm_init_owner (mm, p); +mutex_init (&mm->arg_lock);if (likely (!mm_alloc_pgd (mm))) { mm->def_flags =0 ;kernel/sys.c +22
-0@@ -1564,6 +1564,28 @@ SYSCALL_DEFINE5(prctl, int, option, unsigned long, arg2, unsigned long, arg3, error =0 ;break ; +case PR_SET_PROCTITLE_AREA : { +struct mm_struct *mm = current->mm; +unsigned long addr = arg2; +unsigned long len = arg3; +unsigned long end = arg2 + arg3; + +if (len >PAGE_SIZE ) +return -EINVAL ; + +if (addr >= end) +return -EINVAL ; + +if (!access_ok (VERIFY_READ , addr, len)) +return -EFAULT ; + +mutex_lock (&mm->arg_lock); + mm->arg_start = addr; + mm->arg_end = end; +mutex_unlock (&mm->arg_lock); + +return 0 ; + }default : error = -EINVAL ;break ;Signature · 1 line
1.6.2.5
Signature · 1 line
Live like a child, think like the god.
Andrew Morton
maintainerQuoted 60 lines · On Wed, 4 Nov 2009 00:26:44 +0900 (JST) KOSAKI Motohiro…
========================================
Subject: [PATCH v5] Added
PR_SET_PROCTITLE_AREAoption forprctl()
From: Timo Sirainen <[email protected]>Currently glibc2 doesn't have setproctitle(3), so several userland daemons attempt to emulate it by doing some brutal stack modifications. This works most of the time, but it has problems. For example:
% ps -ef |grep avahi-daemon
avahi 1679 1 0 09:20 ? 00:00:00 avahi-daemon: running [kosadesk.local]# cat /proc/1679/cmdline avahi-daemon: running [kosadesk.local]This looks good, but the process has also overwritten its environment area and made the environ file useless:
# cat /proc/1679/environ adesk.local]Another problem is that the process title length is limited by the size of the environment. Security conscious people try to avoid potential information leaks by clearing most of the environment before running a daemon:
# env - MINIMUM_NEEDED_VAR=foo /path/to/daemonThe resulting environment size may be too small to fit the wanted process titles.
This patch makes it possible for userspace to implement
setproctitle()cleanly. It adds a newPR_SET_PROCTITLE_AREAoption forprctl(), which updates task'smm_struct->arg_startandarg_endto the given area.
test_setproctitle.c
================================================
#include <string.h>
#include <stdlib.h>
#include <unistd.h>
#include <stdio.h>
#include <sys/prctl.h>#define ERR(str) (perror(str), exit(1))
void settitle(char* title){ int err; err = prctl(34, title, strlen(title)+1); if (err < 0) ERR("prctl ");}
void main(void){ long i; char buf[1024]; for (i = 0; i < 10000000000LL; i++){ sprintf(buf, "loooooooooooooooooooooooong string %d",i); settitle(buf); }
Quoted 1 line
}
What happens if userspace unmaps the memory after telling the kernel to use it?
Will processes which try to read the command line get an error reading /proc? If so, do all the commandline-reading programs in the world handle this in an appropriate fashion?
Quoted 72 lines
fs/proc/base.c +22
-9@@ -255,32 +255,45 @@ static int proc_pid_cmdline(struct task_struct *task, char * buffer)int res =0 ;unsigned int len;struct mm_struct *mm =get_task_mm (task); +if (!mm) goto out; + +/* The process was not constructed yet? */ if (!mm->arg_end) goto out_mm;/* Shh! No looking before we're done */ - len = mm->arg_end - mm->arg_start; - +mutex_lock (&mm->arg_lock); + len = mm->arg_end - mm->arg_start;if (len >PAGE_SIZE ) len =PAGE_SIZE ; - + res =access_process_vm (task, mm->arg_start, buffer, len,0 ); +if (mm->arg_end != mm->env_start) +/* prctl(PR_SET_PROCTITLE_AREA) used */ + goto out_unlock; -// If the nul at the end of args has been overwritten, then -// assume application is using setproctitle(3). +/* +* If the nul at the end of args has been overwritten, then assume +* application is using sendmail's SPT_REUSEARGV style argv override. +*/ if (res >0 && buffer[res-1 ] !='\0' && len <PAGE_SIZE ) { len =strnlen (buffer, res); -if (len < res) { - res = len; - }else { +if (len < res) + res = len; +else { len = mm->env_end - mm->env_start;if (len >PAGE_SIZE - res) len =PAGE_SIZE - res; - res +=access_process_vm (task, mm->env_start, buffer+res, len,0 ); + res +=access_process_vm (task, mm->env_start, + buffer+res, len,0 ); res =strnlen (buffer, res); } } + +out_unlock: +mutex_unlock (&mm->arg_lock); + out_mm:mmput (mm); out:include/linux/mm_types.h +1
-0@@ -12,6 +12,7 @@#include <linux/completion.h> #include <linux/cpumask.h> #include <linux/page-debug-flags.h> +#include <linux/mutex.h> #include <asm/page.h> #include <asm/mmu.h>
include/linux/mm_types.h struct mm_struct { · line 243
Quoted 7 lines
include/linux/mm_types.h +1
-0@@ -236,6 +237,7 @@ struct mm_struct {unsigned long stack_vm, reserved_vm, def_flags, nr_ptes;unsigned long start_code, end_code, start_data, end_data;unsigned long start_brk, brk, start_stack; +struct mutex arg_lock;unsigned long arg_start, arg_end, env_start, env_end;unsigned long saved_auxv[AT_VECTOR_SIZE ];/* for /proc/PID/auxv */
Please document the role of arg_lock with a code comment here.
Quoted 36 lines
include/linux/prctl.h +3
-0@@ -90,4 +90,7 @@#define PR_MCE_KILL 33 +/* Set process title memory area for setproctitle() */ +#define PR_SET_PROCTITLE_AREA 34 +#endif /* _LINUX_PRCTL_H */ kernel/fork.c +1
-0@@ -459,6 +459,7 @@ static struct mm_struct * mm_init(struct mm_struct * mm, struct task_struct *p) mm->cached_hole_size = ~0UL ;mm_init_aio (mm);mm_init_owner (mm, p); +mutex_init (&mm->arg_lock);if (likely (!mm_alloc_pgd (mm))) { mm->def_flags =0 ;kernel/sys.c +4
-0@@ -1564,6 +1564,28 @@ SYSCALL_DEFINE5(prctl, int, option, unsigned long, arg2, unsigned long, arg3, error =0 ;break ; +case PR_SET_PROCTITLE_AREA : { +struct mm_struct *mm = current->mm; +unsigned long addr = arg2; +unsigned long len = arg3;
kernel/sys.c SYSCALL_DEFINE5(prctl, int, option, unsigned long, arg2, unsigned long, arg3, · line 1580
Quoted 10 lines
kernel/sys.c +10
-0+unsigned long end = arg2 + arg3; + +if (len >PAGE_SIZE ) +return -EINVAL ; + +if (addr >= end) +return -EINVAL ; + +if (!access_ok (VERIFY_READ , addr, len)) +return -EFAULT ;
It's unobvious (to me) why this access_ok() check is here. If that wasn't totally dumb of me, please add a comment so the next reader won't be similarly mystified.
Quoted 10 lines
kernel/sys.c +7
-0+mutex_lock (&mm->arg_lock); + mm->arg_start = addr; + mm->arg_end = end; +mutex_unlock (&mm->arg_lock); + +return 0 ; + }default : error = -EINVAL ;break ;
Quoted 6 lines · Andrew Morton
What happens if userspace unmaps the memory after telling the kernel to use it?
Will processes which try to read the command line get an error reading /proc? If so, do all the commandline-reading programs in the world handle this in an appropriate fashion?
This case can already occur in the current code; the userspace process would have to munmap() the top of its stack, but it certainly can do so if it tries. In any case, access_process_vm() then returns 0 because of the fault, and thus /proc/pid/cmdline is seen to have zero length. Since a zero-length /proc/pid/cmdline occurs with kernel threads as well, we know this isn't a problem.
2009/11/10 Bryan Donlan <[email protected]>:
Quoted 9 lines · Andrew Morton
Quoted 7 lines
What happens if userspace unmaps the memory after telling the kernel to use it?
Will processes which try to read the command line get an error reading /proc? If so, do all the commandline-reading programs in the world handle this in an appropriate fashion?
Quoted 6 lines
This case can already occur in the current code; the userspace process would have to
munmap()the top of its stack, but it certainly can do so if it tries. In any case,access_process_vm()then returns 0 because of the fault, and thus/proc/pid/cmdlineis seen to have zero length. Since a zero-length/proc/pid/cmdlineoccurs with kernel threads as well, we know this isn't a problem.
Plus, ps can read under exiting process. In this case, task->mm is NULL and proc_pid_cmdline return 0.
procps tools are already NUL safe since long time ago.